Summary Bullets:
• A new digital assistant from Facebook seeks differentiation from Apple Siri, Google Now and others through combined human- and AI-informed actions.
• Likewise, IT buyers considering the application of AI within big data projects should invest in people who know both data and the business itself.
Tucked away within the original run of 60s television series Star Trek (the one with Kirk, Spock and crew), there’s a gem of an episode entitled “City on the Edge of Forever,” which was written by noted science fiction author Harlan Ellison. In that episode, the Enterprise stumbles upon a mysterious, powerful, perhaps even omnipotent machine of sorts called The Guardian of Forever. When Kirk innocently asks his colleagues what it is, the mysterious, mechanistic entity that has the power (as we later discover) to reveal earth’s living past replies:
“A question! Since before your sun burned hot in space and before your race was born, I have awaited a question.”
That line has stuck with me for many years, of course as a reminder of just how insignificant we are in the grand mechanism that is the cosmos. But, more poignantly, it reminds me that we as a species long to remember and be remembered. We learn from the past. We anchor our sense of self to it. And most importantly, we foretell the future through our admittedly malleable recollection of what went before.
In the case of “City on the Edge of Forever,” our fictional characters go so far as to unravel both their present and their future by unwittingly altering their past. That would never have happened if the beings who constructed this guardian had simply posted a human (or some other sentient species) as a security guard to prevent Kirk and crew from doing such an incredibly stupid thing.
And fortunately for us, such shenanigans aren’t likely to happen with Facebook’s newly launched personal digital assistant, Facebook M, because unlike competitive cognitive AI assistants Google Now, Apple Siri and Microsoft Cortana, Facebook’s guardian of forever will be run by both people and machines working in tandem. So far as I can tell, the company’s thinking here is that humans will be able to fill in the cognitive and functional gaps of an AI otherwise left to its own devices. The end result is a system that can actually complete tasks on your behalf, such as booking a table at a restaurant or cancelling your cable service.
As I’ve discussed before (please see “Hello, Cortana. This Is Big Data Calling”, June 16, 2015), I believe these AI-driven personal assistants will not just inform, but also drive the way users interact with data in the enterprise. They are the future of big data. But, they can only go so far. As we’ve seen borne out by IBM’s Watson technologies, true leaps of functionality require smart humans to fill in the gaps. With IBM Watson, that means humans creating (training) smarter data sets by asking Watson questions and grading its responses.
As Facebook progresses with M, tying in data from its Social Graph and improving its cognitive library of knowhow, this avatar will become both smarter and more autonomous, allowing the company to scale back human intervention. That is how all AI-infused enterprise data and analytics projects should proceed. Whether at the grand scale of IBM Watson or with a more humble beginning such as recommendations (chart types, pivot points, transformations, etc.), enterprise buyers must invest in people who know both data and the business itself. In short, someone must protect us from blind faith in our machines; someone must stand guard at the edge of forever.
IT Connection 